Efficient and Privacy-Preserving Federated QoS Prediction for Cloud Services

摘要: With the widespread adoption of cloud computing, large-scale online applications composed services have been deployed in many critical areas. In order to ensure performance applications, Quality Service (QoS) is a key indicator commonly used for service selection and adaptation. Previous studies proposed collaborative QoS prediction approaches estimate personalized values. However, encounters privacy problems practice. As result, threat has become challenge make practical. this paper, we privacy-preserving approach employing federated learning techniques tackle grand challenge. We further improve efficiency by reducing system overhead feasible. The evaluated on real-world dataset, experimental results confirm its effectiveness efficiency.